PREreview of A pangenomics-enabled platform for the high-throughput discovery of antifungal resistance factors in crop pathogens

Reviewer Comments

Puccetti et al. present a pangenomic and GWAS-enabled survey of fungicide resistance in the wheat pathogen Zymoseptoria tritici, analyzing over 1,400 isolates collected across Europe. By integrating high-throughput phenotyping with both reference-based and k-mer–based genotyping approaches, the authors identify resistance-associated variants for more than 40 fungicides spanning multiple chemical classes. The study highlights the roles of SNPs, indels, structural variants, copy number variation, and transposable element insertions in shaping resistance, and further validates novel resistance determinants in known target genes, including SdhC paralogs and beta-tubulin. The breadth of data collected makes this work a valuable contribution to understanding the genetic basis of fungicide resistance, and the resulting resource will be of interest to researchers studying fungal adaptation and crop protection. Importantly, the Croll lab has now set up an extremely powerful pipeline to quickly assay the potential for emergence of resistance against virtually any fungicides.

Overall, this is a well-executed and ambitious study, but there are a number of areas where clarification and refinement would improve readability and reproducibility. Below, we I provide specific comments and suggestions that we I hope will be helpful to the authors.

Introduction

• Lines 90–95: consider rephrasing (e.g., comprehensive → representative; over 40 → a precise number; exhaustive → extensive).

• The authors could briefly explain fungicides and resistance genes in the introduction.

• A summary table of fungicides in the introduction would be helpful.

• Line 103: “relevant geographic and temporal scales” could be rephrased as time and space.

Figure 1

• Figures 2A and 2E, which are summary statistics of the dataset. The similar figure could be added in Figure 1.

• Figure 1D might be better presented as a violin plot.

Figure 2

• Figure 2B is visually too complicated and not easy to see the difference.

• Figures 2D and 2H are interesting, but the y-axis could be adjusted. It would also be better to label which peaks correspond to which genes. Perhaps include a inset “zoom” on the peak of interest in the top right hand quadrant of each graph, to improve clarity.

• Figure 2F shows no significant difference between West and East populations. Why then does the GWAS indicate a peak specifically in the West population?

• For Figure 2F, plotting the original data points on top of the boxplots would be better.

Figure 3

• Figures 3A and 3B: is there any statistical analysis for the plateau? (May not be essential, but worth clarifying.)

• The authors should specify significance levels (p-values).

• Figures 3C and 3D: Is the Y axis the proportion of significant SNPs, rather than genes? With the genes these SNPs map to later indicated via colour?

Figure 4

• Emphasizing significance (p-values) is more important than simply reporting the number of significant genes or SNPs.

• Figure 4B: the Y-axis is labelled as “Frequency strains”. It is now clear what this means. Consider changing it to “Number of strains”.

• How many strains were included in this analysis? Figure 4A shows only 96 strains. Were more strains used but growth plates not pictured here? Please specify.

Figure 5

• The authors could explain the fungicides mentioned either in the introduction or early in the results.

• What does QOI stand for? Please elaborate.

• As mentioned above, gene names for highlighted regions would add clarity. 

• Figure D, bottom Manhattan plot. What is the peak on Chromosome 7? Does it correspond to Cyp51?

Figure 6

• Figure 6A: the colors used (blue for coding, red for non-coding) are difficult to distinguish. Also, consider indicating on the figure the <5% minor allele frequency cutoff mentioned on line 188.

Figure 7

• Figure 7A: inconsistent use of gene names; please standardize.

• Figure 7A: The notation should be consistent: –log10(p) rather than log10(p).

• Figure 7A: the X axis is labelled “Gene names” yet no such names are given.

• Line 208: should read Figure 7A instead of Figure 6A.

• Line 209: consider rephrasing “distinct.”

• Line 210: 61.1% should be corrected to 61.8% as 557/901=0.6182…. Some of the other percentages are also wrong. Also, the terms loci and SNPs are used interchangeably in this paragraph but with a dense genetic map several SNPs can reside in the same gene locus. It would be clearer if “loci” is reserved for when specific gene loci are being discussed.

• Figure 7B: currently unclear. Like in Figure 6C, the percentage of aligned k-mers should be represented as a single score. Why do the scores vary across fungicide treatments? Were only significant k-mers included? This information is missing.

• Figure 7C: does IPO correspond to IPO323? Please clarify.

• Figure 7D: why do annotated genes and TEs appear to show synteny? Are the plots too dense?

Figure 8

• Figure 9 is cited before Figure 8.

• Figure 8D: consider plotting the original data points on violin plots.

Figure 9

• Interesting results. The figure suggests that two different fungicides need to be used together. Are these already combined in practice? If so, the results could be linked more explicitly to agricultural application.

• Figure 9A&B: these non-standard scatterplots do not convey a clear message. If the idea is to quantify the strains’ propensity to only resist one fungicide but not the other, consider using bar & whisker plots.

Figure 10

• Figure 10A: numerous hits throughout the genome surpass the line which presumably indicates statistical significance. Please comment on this and discuss what method you used to decide on the level of significance and whether this was appropriate.

• Figure 10F: what does the white line on colonies represent?

Additional Comments on Text

• Line 106: it is unclear what mapping genetic factors “comprehensively” means. The word should perhaps be avoided.

• Line 126: the word “resistance” appears to have been omitted after “mefentrifluconazole”. The same mistake has also been made elsewhere in the manuscript.

• Can the authors truly call this an atlas? Alternatively, is it more accurate to say they established a method or conducted a large-scale screen?

• Please comment on if/how you plan to make this atlas accessible to the community.

• Line 190: When you say you found “186 significant indels and 901 significant SNPs”, what exactly is meant here by the term “significant”?

• Line 194: Reference 30 should be cited here. Ideally, the “19” pan-genomes should be explicitly mentioned; otherwise, readers may confuse this with the European diversity panel.

• Line 307: appears to have an incomplete citation “()Figure” — please fix.

• Line 346: consider changing “genotype” to “capture” or “analyze”.

• Line 357: change citation format to match other citations

• Line 384: consider deleting the term “comprehensive”.

• Line 502: SNPEFF should be corrected to SnpEff.

• Line 893: should be unaligned rather than un aligned.

• The authors could describe the functions of both Cyp51 and Mfs1.

• Echinocandins: according to the caption of Fig. 7, do they fall under the “Rest” category? Please clarify.

• The effects of fungicide on growth were measured in the context of in vitro growth on media. Could you comment on how good of a proxy this is for the fungicides’ effect on fungal growth in planta?

• Throughout the texts, fungicides are introduced and assayed seemingly on an ad hoc basis. This can be confusing to the reader. It might be worthwhile to introduce every new fungicide when it is first used, what family it is from, as well as explain why the switch was made from the fungicide used in prior assays. Additionally, please include information about the relative use of the fungicides by farmers over time and which fungicides are typically used in combination with one another.

• The authors convincingly explain that the development of fungicide resistance is a major problem (Abstract, line 31-36; Introduction, line 49-53), and how their pipeline facilitates discovery of mechanisms underlying fungicide resistance (Abstract, line 40-43; Introduction, line 90-95). However, it is still unclear how a greater understanding of these mechanisms will ultimately lead to effective use of existing fungicides, or the development of new fungicides that subvert these mechanisms of resistance. Please expand on this point either in the Introduction (line 50-52), or in the Discussion.

  • The authors should emphasize this point especially with regard to data presented in Figure 9. Here, it is clearly demonstrated that Z. tritici strains are either resistant to diethofencarb or carbendazim. Therefore, for successful eradication of an unknown Z. tritici strain in the field, farmers should spray a mixture of both diethofencarb and carbendazim. This is a real-world application of the authors’ advances in understanding B-tubulin fungicide resistance, and should be emphasized much more in the Discussion section.

• Are the Z. tritici strains from the European diversity panel equally virulent on wheat? (The panel is mentioned first in line 107-109, and used throughout the manuscript). Some strains of the European diversity panel seem to consistently grow more or less rapidly than others on media (Figure 4A, control images).

• Is the effect of fungicides on other fungi like arbuscule mycorrhizal fungi a consideration during antifungal development or usage?

Competing interests

The authors declare that they have no competing interests.